MPIDR Working Paper
Sibship size and the genetic prediction of education
MPIDR Working Paper WP-2026-034, 67 pages.
Rostock, Max Planck Institute for Demographic Research (July 2026)
Abstract
Sibship size is a widely studied predictor of educational attainment. However, existing research documents its average effect on schooling, not whether it reshapes how strongly children's individual potential translates into attainment. This study tests that question using a child's educational attainment polygenic index (EA PGI) as a marker of educational propensity. Controlling for parental genotype leaves any remaining variation in children's genetic marker randomized, providing a natural experiment that identifies direct genetic effects on completed education. Across more than 17,000 genotyped respondents from two large-scale surveys conducted in the United States, Add Health and the Wisconsin Longitudinal Study, each with parental EA PGIs Mendelian-imputed from genotyped sibling pairs, the EA PGI's effect on years of schooling declines by roughly 20-25 percent from one to six siblings (1.08-0.86 years per SD in Add Health; 0.90-0.68 in WLS). The attenuation is replicated in both cohorts and remains statistically significant in the WLS after conditioning on parental EA PGIs; a related analysis using a parental fertility PGI as a proxy for sibship size supports these findings. Because larger sibships are concentrated among disadvantaged families, this pattern is consistent with resource dilution, and represents a channel through which stratified fertility reproduces educational inequality.
Keywords: USA, children, education, genetics, siblings